Andhra Pradesh: 2
Again, vast swaths of the center-north are missing. The lack of Gujaratis isn’t as problematic because the HapMap has them. The HGDP has Pakistani populations, though Zach L. will be gratified that Zack A. is finally doing some more work on Punabis as a whole (not farmers from a particular community). The SVGP has 80 Singapore Indians, who are presumably going to be mostly Tamil. Behar et al. has two low caste groups from Kerala and Tamil Nadu respectively, as well as another group from northern Karnataka which is not specified (as well as Kerala Jewish samples). Finally, the Xing et al. data set has several South Indian groups, as well as Punjabi Arain. Unfortunately due to its peculiar SNP coverage it doesn’t intersect much with the other samples, so it won’t be used much. But note: the massive undersampling of the center-north and eastern India in all of these public data sets. Additionally, it being South Asia, geographic coverage is probably not enough. There is almost certainly a great deal of genetic difference between a tribe in Kerala and the Nasranis.
So here’s the link to 23andMe. It will cost you $260 over the whole year.